Calculating Enzyme Inhibition with Random Forests.

Area of Science:

• Computational chemistry and cheminformatics
• Artificial intelligence in drug discovery
• Machine learning for molecular modeling

Background:

• Cyclin-dependent kinase 2 (CDK2) is a crucial target for anticancer drug development due to its role in cell cycle regulation.
• Accurate prediction of binding affinity is essential for identifying potent drug candidates.
• Molecular docking simulations provide structural insights into drug-target interactions.

Purpose of the Study:

• To develop a regression model for predicting binding affinity using docked structures.
• To apply the Random Forest machine learning algorithm for enhanced predictive accuracy.
• To create a computational workflow for drug discovery targeting CDK2.

Main Methods:

• Utilized the Random Forest algorithm from the Scikit-Learn library for regression modeling.
• Employed molecular docking results generated by Molegro Virtual Docker (MVD).
• Developed a Google Colab workflow integrating docking data and machine learning models (MVD4ML and SKReg4Model).

Main Results:

• Successfully built a Random Forest regression model to predict the inhibition of cyclin-dependent kinase 2.
• The workflow demonstrated the feasibility of using docking results for generating predictive regression models.
• Code and datasets are publicly available on GitHub for reproducibility and further research.

Conclusions:

• Random Forest regression is a powerful tool for predicting binding affinity in drug discovery.
• The developed workflow streamlines the process of identifying potential CDK2 inhibitors.
• Open-source availability of tools and data promotes collaborative research in computational drug design.